步态识别新视角：GaitSet算法

需积分: 48 20 浏览量更新于2024-09-07 7 收藏 486KB PDF 举报

"步态识别是一种基于人体行走方式的生物特征识别技术，广泛应用于犯罪预防、法医学鉴定和社会安全等领域。由于其能够在远距离下进行识别，因此具有很高的实用价值。在传统的步态识别方法中，通常使用步态模板或步态序列来表示个体的行走特征。然而，步态模板难以保留时间信息，而步态序列则受到严格的顺序约束，限制了识别的灵活性。在给定的研究中，提出了一种新的视角——GaitSet，将步态视为由独立帧组成的集合。这种方法的优势在于对帧的排列顺序不敏感，同时能够自然地整合来自不同场景（如不同视角、穿着不同衣物或携带物品）的视频中的帧。GaitSet网络设计用于从这个帧集合中学习身份信息，即使在帧顺序变化的情况下也能保持识别性能的稳定。 GaitSet网络的结构可能包括多个层次，每个层次通过深度学习技术来提取和融合帧中的关键信息。首先，可能会有预处理步骤，如使用背景 subtraction 技术将行人从复杂的视频背景中分离出来，生成如CASIA-B数据库所示的黑白轮廓图（Silhouette）。接着，这些轮廓图会被输入到GaitSet网络中，网络通过一系列卷积层、池化层和全连接层来捕获帧间的时空特征。由于步态集合的特性，网络设计可能包含注意力机制，以强调对识别身份至关重要的帧或特征。实验结果显示，在正常行走条件下，GaitSet方法相比于传统方法表现出了更高的识别准确率。此外，它还能有效地处理跨视图识别问题，即便个体的衣物、携带物品或观察角度发生变化，也能保持良好的识别性能。这表明GaitSet对于应对真实世界中的挑战，如光照变化、遮挡等因素，具有很强的鲁棒性。 GaitSet为步态识别提供了一个创新的解决方案，它不仅解决了传统方法中的时间和顺序问题，还增强了模型对不同环境条件的适应性。这种方法的出现，无疑将进一步推动步态识别技术在安全监控、智能安防等领域的应用。"

GaitSet: Regarding Gait as a Set for Cross-View Gait Recognition

Hanqing Chao

†

, Yiwei He

†

, Junping Zhang

‡

, JianFeng Feng

Fudan University, Shanghai, China

{hqchao16, heyw15, jpzhang, jffeng}@fudan.edu.cn

Abstract

As a unique biometric feature that can be recognized at a

distance, gait has broad applications in crime prevention,

forensic identiﬁcation and social security. To portray a gait,

existing gait recognition methods utilize either a gait tem-

plate, where temporal information is hard to preserve, or a

gait sequence, which must keep unnecessary sequential con-

straints and thus loses the ﬂexibility of gait recognition. In

this paper we present a novel perspective, where a gait is

regarded as a set consisting of independent frames. We pro-

pose a new network named GaitSet to learn identity informa-

tion from the set. Based on the set perspective, our method

is immune to permutation of frames, and can naturally inte-

grate frames from different videos which have been ﬁlmed

under different scenarios, such as diverse viewing angles,

different clothes/carrying conditions. Experiments show that

under normal walking conditions, our single-model method

achieves an average rank-1 accuracy of 95.0% on the CASIA-

B gait dataset and an 87.1% accuracy on the OU-MVLP gait

dataset. These results represent new state-of-the-art recog-

nition accuracy. On various complex scenarios, our model

exhibits a signiﬁcant level of robustness. It achieves accura-

cies of 87.2% and 70.4% on CASIA-B under bag-carrying and

coat-wearing walking conditions, respectively. These outper-

form the existing best methods by a large margin. The method

presented can also achieve a satisfactory accuracy with a small

number of frames in a test sample, e.g., 82.5% on CASIA-

B with only 7 frames. The source code has been released at

https://github.com/AbnerHqC/GaitSet.

1 Introduction

Unlike other biometrics such as face, ﬁngerprint and iris,

gait is a unique biometric feature that can be recognized at a

distance without the cooperation of subjects and intrusion to

them. Therefore, it has broad applications in crime prevention,

forensic identiﬁcation and social security.

However, gait recognition suffers from exterior factors

such as the subject’s walking speed, dressing and carrying

This work is supported in part by National Natural Science

Foundation of China (NSFC) (Grant No. 61673118) and in part by

Shanghai Pujiang Program (Grant No. 16PJD009).

†

H.C. and Y.H. are co-ﬁrst authors.

‡

Corresponding author.

2019, Association for the Advancement of Artiﬁcial

Figure 1: From top-left to bottom-right are silhouettes of a

completed period of a subject in CASIA-B gait dataset.

condition, and the camera’s viewpoint and frame rate. There

are two main ways to identify gait in literature, i.e., regarding

gait as an image and regarding gait as a video sequence.

The ﬁrst category compresses all gait silhouettes into one

image, or gait template for gait recognition (He et al

2019;

Takemura et al

2018a; Wu et al

2017; Hu et al

2013).

Simple and easy to implement, gait template easily loses

temporal and ﬁne-grained spatial information. Differently,

the second category extracts features directly from the orig-

inal gait silhouette sequences in recent years (Liao et al

2017;

Wolf, Babaee, and Rigoll 2016). However, these methods are

vulnerable to exterior factors. Further, deep neural networks

like 3D-CNN for extracting sequential information are harder

to train than those using a single template like Gait Energy

Image (GEI) (Han and Bhanu 2006).

To solve these problems, we present a novel perspective

which regards gait as a set of gait silhouettes. As a periodic

motion, gait can be represented by a single period. In a sil-

houette sequence containing one gait period, it was observed

that the silhouette in each

position

has unique appearance,

as shown in Fig. 1. Even if these silhouettes are shufﬂed,

it is not difﬁcult to rearrange them into correct order only

by observing the appearance of them. Thus, we assume the

appearance of a silhouette has contained its

position

informa-

tion. With this assumption, order information of gait sequence

is not necessary and we can directly regard gait as a set to

extract temporal information. We propose an end-to-end deep

learning model called GaitSet whose scheme is shown in

Fig. 2. The input of our model is a set of gait silhouettes.

First, a CNN is used to extract frame-level features from

each silhouette independently. Second, an operation called

Set Pooling is used to aggregate frame-level features into

a single set-level feature. Since this operation is applied on

high-level feature maps instead of the original silhouettes, it

arXiv:1811.06186v2 [cs.CV] 18 Nov 2018

下载后可阅读完整内容，剩余7页未读，立即下载

moxiaoguai1993_

粉丝: 1
资源: 9

步态识别新视角：GaitSet算法

论文研究-基于K-近邻算法的人体步态识别 .pdf

步态识别源码

通过GEI子空间投影和协同表示分类进行步态识别

步态识别-基于图卷积网络+人体骨架实现的步态识别算法-附项目源码-优质项目分享.zip

实现人体步态检测和基于步态的人体身份识别算法，主要内容包括背景的提取，人体运动区域检测，人体轮廓提取，步态特征的提取，相似性度量和分类识别。-

基于肢体长度参数的多视角步态识别算法

步态识别-基于特征点加速度的步态识别算法实现-Matlab-附项目源码+流程教程-优质项目分享.zip

基于深度学习的步态识别算法优化研究.pdf

基于广义主成分分析的步态识别算法研究 (2009年)

基于频谱分析的Procrustes统计步态识别算法 (2011年)

最新资源